Mammography Screening

Truth, Lies and Controversy

Peter Gotzsche

• Publisher: CRC Press
• Year: 2021
• ISBN:
  9781000477092
• Edition: 1
• Topic: Medicine
• Subtopic: Oncology

'This book gives plenty of examples of ad hominem attacks, intimidation, slander, threats of litigation, deception, dishonesty, lies and other violations of good scientific practice. For some years I kept a folder labeled Dishonesty in breast cancer screening on top of my filing cabinet, storing articles and letters to the editor that contained statements I knew were dishonest. Eventually I gave up on the idea of writing a paper about this collection, as the number of examples quickly exceeded what could be contained in a single article.' From the Introduction The most effective way to decrease women's risk of becoming a breast cancer patient is to avoid attending screening. Mammography screening is one of the greatest controversies in healthcare, and the extent to which some scientists have sacrificed sound scientific principles in order to arrive at politically acceptable results in their research is extraordinary. In contrast, neutral observers increasingly find that the benefit has been much oversold and that the harms are much greater than previously believed. This groundbreaking book takes an evidence-based, critical look at the scientific disputes and the information provided to women by governments and cancer charities. It also explains why mammography screening is unlikely to be effective today. All health professionals and members of the public will find these revelations disturbingly illuminating. It will radically transform the way healthcare policy makers view mammography screening in the future. 'If Peter Gotzsche did not exist, there would be a need to invent him...It may still take time for the limitations and harms of screening to be properly acknowledged and for women to be enabled to make adequately informed decisions. When this happens, it will be almost entirely due to the intellectual rigour and determination of Peter Gotzsche.' From the Foreword by Iona Heath, President, RCGP 'If you care about breast cancer, and we all should, you must read this book. Breast cancer is complex and we cannot afford to rely on the popular media, or on information from marketing campaigns from those who are invested in screening. We need to question and to understand. The story that Peter tells matters very much.' From the Foreword by Fran Visco, President, National Breast Cancer Coalition

1

Introduction

SOMETIMES YOU CAN’T BELIEVE THAT WHAT YOU’VE JUST SEEN IS actually true. And sometimes it isn’t. An example of this could be if you saw a magician suspend a person in the air without support. However, what I describe in this book is true, although much of it goes far beyond what most people would think scientists were capable of doing, not only to each other but also, more importantly, to the public they are supposed to serve.

Like other people, scientists are often driven by emotions, career aspirations, strong beliefs, money and fame rather than facts and logic. This is not unexpected. However, when it comes to mammography screening, the extent to which some scientists are ready to deny what they see – and here I am not talking about magic – and sacrifice sound scientific principles in order to arrive at politically acceptable results in their research is astounding. I have a vague hope that it might be better in other disciplines, but I doubt it, as it all comes down to human psychology.

This book gives plenty of examples of ad hominem attacks, intimidation, slander, threats of litigation, deception, dishonesty, lies and other violations of good scientific practice. For some years I kept a folder labelled Dishonesty in breast cancer screening on top of my filing cabinet, storing articles and letters to the editor that contained statements I knew were dishonest. Eventually I gave up on the idea of writing a paper about this collection, as the number of examples quickly exceeded what could be contained in a single article.

In November 2009, I attended a meeting in London on cancer screening. The exaggerated claims about mammography screening that I heard at this meeting triggered something for me. Retired breast surgeon Michael Baum also attended the meeting and, finding ourselves thinking along similar lines, we agreed to write a book together. Once having started writing – actually in the middle of the meeting – I just couldn’t stop. I also realised it would be much easier to write the book alone, which Baum happily accepted. He published his own book, Breast beating: A personal odyssey in the quest for an understanding of breast cancer, the meaning of life and other easy questions, on screening in 2010, which is also an autobiography.

Sweden has been a major player in research on mammography screening, encouraging the implementation of screening all over the world. Much of the brouhaha our research on screening has created has occurred in Sweden and Denmark. I believe an account of the events in these countries has wider interest and, as few people can read Swedish and Danish, this book describes these events in English. The book also describes research and events in other countries, with an emphasis on the United States, the United Kingdom and Australia.

My interest in research misconduct began early. I studied biology and chemistry at the universities of Copenhagen, Uppsala and Lund and I acquired a full academic education both in Denmark and in Sweden. Uppsala is the home town of Carl von Linné, the famous taxonomist who gave Latin names to so many of our plants and animals, and the surroundings of the city were ideal for bicycle excursions into the wilderness.

A newly qualified biologist, I was employed by the Swedish drug company Astra in Copenhagen in 1975. In the beginning, I worked as a drug representative and product manager, but I quickly moved into clinical research and established a medical department at a new joint venture company, Astra-Syntex, where I handled regulatory affairs and was responsible for clinical trials.

I knew a good deal about insects, which was my speciality in biology, but virtually nothing about drugs. However, it didn’t take long before I learned there were also plenty of bugs in drug research. I was relatively free to work as I wished at Astra-Syntex and I wasn’t forced to do things against my principles, but I knew this luck couldn’t last forever. The working practices of the drug industry, both in research and in marketing, are problematic, and the industry’s priority is profit-making, which often leads to untoward consequences for the patients and for our national economies. Therefore, I wanted to leave. As I was attracted to the challenges in healthcare, I decided that, rather than going back to biology, where there were few career options, I would study medicine.

I asked for my director’s permission to attend courses in 10% of my full working time, which later became 33% and then increased to 50% during the final years. During this time, I was offered an attractive job as clinical research director for the Nordic area, located at the headquarters of Astra in Sweden, which would have earned me the largest and most well-equipped Volvo that money can buy, and a corresponding salary, but I preferred to finish my studies. It is surprising how effective one can be when there is little time. I finished medical studies in 6½ years, in 1984, left the company and started clinical work at hospitals in Copenhagen. Then came two revelations that influenced me greatly.

The first occurred right after my final exams, during my first appointment. The head of endocrinology suggested my thesis subject should be insulin antibodies, and I collected about 500 scientific articles on the subject. However, I couldn’t make much sense of them. I tried to construct dose-response curves on the effects of insulin using published data, but some of the papers were so incompatible that I realised that not all researchers had been entirely honest.

The next revelation came a year later when I conducted a study where I had carefully worked out the dose of insulin for a physiological experiment, based on a graph from a paper published in the prestigious New England Journal of Medicine. I studied the cardiovascular effects of insulin, and I thought it wouldn’t be possible for any of our diabetes patients to get a dose that was too high. It was pretty shocking, therefore, that the very first patient developed symptoms of hypoglycaemia to such an extent that I needed to give a rescue dose of intravenous glucose. When an endocrinologist external to our research group heard who the first author was of the paper I had trusted, he asked me if I didn’t know him. I didn’t. The endocrinologist explained that the graphs this person published were too good to be true. Thus, it seemed that my patient and I had fallen victims to fraud.

I published our study¹ but I wasn’t proud of it. It had eight patients, but also eight authors, which was a bit much considering only three people did the research. Our study wasn’t particularly useful and I wondered whether we should refrain from asking patients to participate in such studies.

The two revelations about insulin research indicated that not only were serious manipulations with the data common in the drug industry but also that they occurred in research led by academics. I shall never forget how happy I was when I dumped the 500 insulin papers in the rubbish. Goodbye insulin and goodbye endocrinology. I decided to turn my negative experience into a positive one by embarking on a very different project for my thesis: bias in clinical research.

I also decided to work alone. I wasn’t happy with the prevailing tradition of having many authors on papers that only few people had contributed to, often including the chair of the department for no other reason than being the chair of the department. Some doctors’ creative ideas about authorship made one of my colleagues remark that if a doctor had lent Shakespeare a pencil, he would have become co-author of Macbeth. And it led to an amusing letter in the New England Journal of Medicine with the title ‘Et al. gets Nobel Prize’² (when there are many authors on a paper, it is customary to mention only the first few and to group the rest as ‘et al.’, which means ‘and others’). I was inspired by Charles Darwin, who had written On the Origin of Species alone, which I consider the most important piece of research that has ever been done, and I was also inspired by a relative of mine who was sole author of the six papers that comprised his thesis.

I wrote a protocol for the thesis and assembled 244 reports of trials that had compared one anti-arthritis drug with another. The thesis, ‘Bias in double-blind trials’,³ which was based on six papers,^{4, 5, 6, 7, 8, 9} was defended in 1990. It was the first time a whole therapeutic area had been so thoroughly investigated, and the thesis changed my career. The ink was hardly dry before UK statistician Douglas Altman picked it up at the library, read, and sent it to the director of the National Perinatal Epidemiology Unit, Iain Chalmers, in Oxford, urging him to read it. I didn’t know who Iain Chalmers was, but my ignorance didn’t last long. He invited me to Oxford in 1992 when the British minister of health opened the UK Cochrane Centre. I opened The Nordic Cochrane Centre in Copenhagen in 1993 and abandoned clinical work in 1997.

Iain Chalmers started The Cochrane Collaboration in 1993 and was later knighted for his outstanding achievements. It built on a common frustration among researchers and others that much medical research was of poor quality and was biased, and on the realisation that we lacked updated, reliable systematic reviews of the randomised trials that could tell us what the benefits and harms of our interventions were. We couldn’t offer our patients the best treatments, as the medical research literature was such a mess that we didn’t know which treatments were best.

Once established, The Cochrane Collaboration caught on and grew quickly. It is a registered charity and represents the state of the art for the conduct and dissemination of systematic reviews, as described in the 649-page Cochrane Handbook.¹⁰ The collaboration currently involves more than 25 000 people. The reviews are published electronically in The Cochrane Library, and there are more than 4000 such reviews, which are regularly updated. Half of the world’s population have free access to the full reviews through national subscriptions financed by their national governments; the other half have access to the abstracts.

My thesis had other ramifications. It led to an invitation by statistician David Moher in Ottawa to attend a workshop in 1993 aimed at defining criteria for scoring the reliability of randomised trials. A few hours into the workshop, the late Thomas C Chalmers suggested that it was futile to try to agree on a quality score for assessing trials. It wouldn’t work and we should rather focus on the various biases, one by one. All agreed to his proposal, and the agenda was changed into one of suggesting a set of guidelines for transparent reporting of trials that enabled the readers to assess their reliability. I co-authored the CONSORT guidelines for good reporting of randomised trials that most top journals refer to in their instructions to authors, as well as co-authoring similar guidelines for observational studies (STROBE) and systematic reviews (PRISMA). Currently, the research groups I am a member of work on guidelines for writing protocols for trials (SPIRIT) and protocols for systematic reviews (PRISMAP).

My thesis focused on statistical issues. What I particularly love about mathematics is that, in contrast to medicine, where one has to learn so much by heart, it isn’t necessary to remember much, as one can deduce what is needed based on rather few premises. At the exam in mathematics at the university, I drew a question, started in the upper left corner of the green-board and ended in the lower right corner after having filled it entirely with formulae. The only comment from my examiners was ‘Thank you’. They gave me the highest mark on an elaborate scale where this mark was used only in exceptional cases.

The reason I tell you this is to provide a little balance. My opponents in the scientific debates about mammography screening have often accused me of having made numerous errors in my calculations. As far as I know, I haven’t made any, and I shall show in this book that those accusations that were concrete, and not just mere utterings and condemnations, were wrong. I shall also explain why so many of the analyses performed by professors o...